US 8004278 B2
Probes are electrically connected to a surface of a tunnel junction film stack comprising a free layer, a tunnel barrier, and a pinned layer. Resistances are determined for a variety of probe spacings and for a number of magnetizations of one of the layers of the stack. The probe spacings are a distance from a length scale, which is related to the Resistance-Area (RA) product of the tunnel junction film stack. Spacings from as small as possible to about 40 times the length scale are used. Beneficially, the smallest spacing between probes used during a resistance measurement is under 100 microns. A measured in-plane MagnetoResistance (MR) curve is determined from the “high” and “low” resistances that occur at the two magnetizations of this layer. The RA product, resistances per square of the free and pinned layers, and perpendicular MR are determined through curve fitting.
1. A method for determining at least a portion of a hysteresis curve of an unpatterned tunnel junction film stack having a surface, comprising the steps of:
placing a multi-point probe comprising a plurality of probes in electrical contact with the surface of the unpatterned tunnel junction film stack, wherein the smallest spacing between any two of the probes is less than 100 microns; and
measuring, using the multi-point probe, resistance as a function of applied magnetic field, the step of measuring determining at least a portion of the hysteresis curve.
2. The method of
3. The method of
The instant application is a divisional of U.S. patent application Ser. No. 11/119,665, filed May 2, 2005, which is a divisional of U.S. patent application Ser. No. 10/244,766, filed Sep. 16, 2002, each incorporated by reference herein.
This invention was made with Government support under grant contract number MDA972-99-C-0009 awarded by the Defense Advanced Research Projects Agency (DARPA) of the United States Department of Defense. The Government has certain rights in this invention.
This invention relates to tunnel junction devices and, more specifically, the invention relates to electrically characterizing tunnel junction film stacks.
Conventional techniques for characterizing tunnel junction film stacks require extensive processing of a semiconductor wafer in order to measure a few characteristics of the tunnel junction film stack. This processing is time consuming, complex, and can ruin devices created via processing. Additionally, even if processing creates suitable devices for test, it is unclear whether measurements have been influenced by the processing. In other words, it is unclear as to whether the measurements are a function of the tunnel junction film stack, the processing that creates additional structures needed to measure characteristics of the tunnel junction film stack, the additional structures, or some combination of these.
Consequently, a need exists for characterizing tunnel junction films without time consuming and expensive processing that is fraught with accompanying flaws.
Aspects of the present invention overcome problems of the prior art by providing techniques to characterize tunnel junction film stacks with minor or no processing. In one aspect of the invention, resistances are determined for a variety of probe spacings. Probes are electrically connected to a surface of a tunnel junction film stack. Generally, a tunnel junction film stack comprises a free layer, a tunnel barrier, and a pinned layer, but many other configurations are possible. The electrical connection can be a physical connection directly to the surface of the tunnel junction film stack or an electrical connection to a contact pad that is physically attached to the surface of the tunnel junction film stack. The probe spacings are generally selected to be within a predetermined distance from a length scale, which is related to the Resistance-Area (RA) product of the tunnel junction film stack. While other spacings may be used, resistances, depending on probe or contact pad configurations, tend to stabilize after probe or contact pad spacing reaches a certain distance from the length scale. For instance, probe spacings greater than 40 times the length scale generally yield very similar resistances. The resistances are measured for both magnetizations of one of the layers of the stack, such that sets of “high” and “low” resistances are determined for both magnetizations of the free layer. Generally, the free layer is the layer able to have multiple magnetizations. Additionally, an in-plane MagnetoResistance (MR) curve can be determined from the high and low resistances that occur at the two multiple magnetizations of the free layer.
In another aspect of the invention, the RA product, perpendicular MR, and resistances per square of the free layer, RF, and of the pinned layer, RP, are determined through a curve fitting technique. While a number of curve fitting techniques may be used to determine these quantities, a suitable technique is as follows. The in-plane MR is determined from the measured high and low resistances using a standard formula. The in-plane MR and one or both of the resistances are simultaneously fit to theoretical values of in-plane MR and resistance. This is done by assuming values for the RA product, RF, RP, and perpendicular MR and then calculating in-plane MR and resistance curves. The values of the in-plane MR and resistance curves are then compared with the calculated values of in-plane MR and resistance curves. This process is then iterated, each time changing values for the RA product, RF, RP, and perpendicular MR, until the best agreement between the measured and calculated values of in-plane MR and resistance is obtained.
In another aspect of the invention, an apparatus is provided for characterizing tunnel junction film stacks. The apparatus comprises a magnetic field generator used to generate a magnetic field, a micro-machined multi-point probe having four or more probes, where the smallest spacing between any two of the multiple probes used during a resistance measurement has a spacing of 100 microns or less, and a resistance measuring module coupled to the multi-point probe and adapted to measure resistance. The magnetic field is generated to place a semiconductor wafer having a tunnel junction film stack into one of a plurality of magnetizations for the tunnel junction film stack. A resistance measurement by the resistance module at this magnetization at least partially characterizes the tunnel junction film stack. Additionally, various probes and contact pad configurations are described that are suitable for use with the apparatus. Advantageously, a multi-point probe is described that allows many different voltage measurements to be taken very quickly. When using the multi-point probe, the apparatus further comprises a multiplexer used to couple probes to the resistance measuring module. Probe spacings are generally selected to be within a predetermined distance from a length scale, which is related to the RA product of a tunnel junction film stack being measured.
A more complete understanding of the present invention, as well as further features and advantages of the present invention, will be obtained by reference to the following detailed description and drawings.
The present invention overcomes limitations of conventional techniques because little or no processing of a semiconductor wafer having a tunnel junction film stack formed thereon is necessary to characterize the tunnel junction film stack. The present invention uses probes or contact pads in particular configurations to characterize tunnel junction film stacks. Resistance is determined by using the probes, contact pads, or both, and by using current and voltage through known resistance-measuring techniques. The resistance is measured for a particular magnetization of a free layer and certain spacing or spacings between probes or contact pads. The free layer is magnetized with a different magnetization, and another resistance is determined. One of the previously determined resistances is a high resistance, and one is a low resistance. Using these two determined resistances, an in-plane MagnetoResistance (MR) may be determined. This process is repeated for a number of probe or contact pad spacings. An in-plane MR curve and two resistance curves, one for each of the high and low resistances, can be determined from the measured data. Using these curves, values are determined through curve fitting techniques for resistance per square of the free layer (RF), resistance per square of the pinned layer (RP), the Resistance-Area (RA) product of the stack, and the perpendicular MR.
For ease of reference, this disclosure is divided into three sections: (1) the INTRODUCTION section describes why commonly held beliefs about the inability to characterize tunnel junction stacks by using probes or contact pads are incorrect, and gives a simple description of aspects of the present invention; (2) the EXEMPLARY THEORY AND METHOD section describes a method suitable for characterizing tunnel junction stacks with probes or contact pads and describes theory that explains measured results; and (3) the EXEMPLARY APPARATUS section describes apparatus suitable for use with the present invention.
A tunnel junction film stack is a series of films formed on top of each other. A tunnel junction film stack is used to create a tunnel junction device, to be described in more detail below. An unprocessed semiconductor wafer 100 is shown in
A tunnel junction is a device where carriers, generally electrons, are made to “tunnel” through a tunnel barrier. Diagrams of a simple tunnel junction device 106 are shown in
The pinned layer 150 is named as such because its magnetization 170 is pinned such that it “points” in a certain direction, as shown in
Because the current is known, the voltage between where the current enters and where it leaves provides an indication of the resistance between the entry and exit points. This resistance changes depending on magnetizations 160, 170: the parallel magnetization, shown in
The resistance measured by current that passes perpendicular, as illustrated by reference 190, to the surface 135 of an unprocessed stack 110 or processed stack 111 is quantified by a Resistance-Area (RA) product. The RA product changes when the magnetization 160 is changed.
Thus, there are multiple quantities that a researcher would like to characterize in order to determine the best materials, thicknesses, and processing procedures for the unprocessed stack 110, as shown in
Even when conventional techniques are used for creating tunnel junction devices suitable for measuring the RA product, the resistance per square of the free layer and pinned layer, RF and RP, respectively, cannot be determined. The resistance per square of a layer is the resistance a current experiences as it flows through the layer, parallel to upper and lower surfaces of the layer. The resistances RF and RP are quantities that it would be beneficial to measure but that cannot be measured through conventional techniques used to measure the RA product.
Moreover, it was previously believed that MR and other quantities could not be measured by using probes or contact pads in electrical connection with the surface of a tunnel junction film of a tunnel junction film stack (generally called a “stack” herein for simplicity). A tunnel junction film stack is any stack having one or more tunnel barriers sandwiched between two or more other films. In this disclosure, a “probe” is a device that is not affixed to the surface being measured, while a “contact pad” is affixed to the surface being measured. A probe can be affixed to an intermediate metal structure, such as a contact pad. Electric coupling of the probe to the tunnel junction film stack can occur through physical contact between the probe and the surface of the tunnel junction film stack, through a probe that is affixed to an intermediate metal structure, or through other techniques known to those skilled in the art.
The four-point probe 300 is a well known tool used to characterize many different types of semiconductors and other materials. In fact, it is used to characterize Giant MagnetoResistive (GMR) films, which are used in read and write heads of many current hard drives. In GMR, the tunnel barrier 140 is replaced by a metal, which allows current to flow to the pinned layer 150.
However, for tunnel junction devices, conventional thought was that the current could not pass through the tunnel barrier 140. The current, it was assumed, passed through path 350 and never into the pinned layer 150. Consequently, the RA product of the tunnel junction barrier 140 could not be determined, as the current never passes through pinned layer 150. Because of this, the MR cannot be determined, as current must flow through tunnel junction barrier 140 in order to determine MR. Moreover, when a four-point probe 300 was used to attempt to measure MR, the measured MR would be very low, as shown below. Ironically this was because the current flowed through both the free and pinned layers (and so did pass through the barrier) but the contribution of the barrier to the resistance was negligibley small. Therefore, any researcher using commonly available probes (such as four-point probe 300) to attempt to characterize tunnel junction films would determine no suitable data. Additionally, while resistance can be measured with a four-point probe 300, it was previously believed that the measured resistance was not useful.
What the researchers did not realize, and what is part of the present invention, is that the optimum probe spacing is related to the RA product and the resistance per square of the free and pinned layers (RF and RP, respectively) of the unprocessed stack 110. As developed more fully below, spacing between probes must be near a length scale, defined herein as follows:
The theory behind aspects of the present invention is described below. Before describing theory, it is beneficial to describe a rudimentary model that comes close to theory and to show some figures that illustrate why conventional probes do not measure suitable quantities for tunnel junction film stacks.
Referring now to
While a predetermined current is forced through the contact pads 410, 420, voltage is measured. The measured voltage allows a measured resistance to be determined for the particular contact spacing, L. This step measures one of RH or RL, depending on the magnetization of the free layer 130. A magnetic field is applied to cause the magnetization of the free layer 130 to point in a different direction, then current is again forced through the contact pads 410, 420, and the resultant voltage is measured. Because RH and RL are known, an in-plane MR may be determined. This process is repeated for many different contact spacings, L. Measured resistances and in-plane MR may be plotted, as shown in
It should be noted that the curves in
Nonetheless, these figures also show that characterization of tunnel junction stacks may be made easily and with little or no processing, as long as the choice of probe or contact pad spacing or spacings is correctly chosen.
Exemplary Theory and Method
The graphs in
Theory for Cylindrical Symmetry
Consider the case of cylindrical symmetry (for example, assume a single point source of current at the origin). The equations are as follows:
Combining these three equations gives the following differential equation:
For the case of four equally spaced probes in a line, the solution is as follows:
Consider the case of four probes in a straight line in the order of I+, V+, V−, and I−, where spacings between the probes are as follows: a, which corresponds to the spacing between I+ and V+; b, which corresponds to the spacing between V+ and V−; and c, which corresponds to the spacing between V− and I−. For this situation, the solution is as follows:
Consider the case of four probes placed arbitrarily on the surface of a wafer (not necessarily in a straight line), where spacings between the probes are as follows: α, which corresponds to the spacing between I+ and V+; b, which corresponds to the spacing between I+ and V−; c, which corresponds to the spacing between V+ and I−; and; d, which corresponds to the spacing between V− and I−. For this situation, the solution is as follows:
As a second example, consider the case of linear symmetry (for example,
For the specific case shown in
Now that the theory has been explained for a variety of probe and contact pad configurations, it will now be shown how to use this theory in order to characterize tunnel junction film stacks.
Referring now to
Method 700 begins in step 710, when a number of resistances are measured for the high and low resistances. The high and low resistances are caused by first and second magnetizations of the free layer (or other layer able to change magnetizations) of the stack. These resistances are measured for a variety of probe or contact pad spacings. As explained above, it is beneficial to measure resistance by using spacings from the smallest possible spacing to about 40λ. The “spacing” that is important depends on the probe or contact pad configuration being used. For instance, for a four-point probe the spacing that is important is the average of the three distances between the four probes in the four-point probe. Regardless of the probe or contact pad configuration used, a number of probe spacings should be used in order to determine enough resistance data to adequately characterize the tunnel junction film stack.
There are a number of techniques for performing step 710. For instance, when using four-point probes, a selected four-point probe, having a selected spacing between probes, is placed in electrical contact with a surface of the stack (e.g., the surface of the free layer), then current is passed through two of the probes while the voltage is measured at another two of the probes. It should be noted that voltage could be applied to the stack and current measured. A magnetic field, suitable to switch magnetization of the free layer, is applied to the stack. Then the current is again passed through two of the probes while voltage is measured at another two of the probes. Another four-point probe, with different probe spacing, is selected and measurements are taken at both magnetizations of the free layer. This process is repeated until all spacings have been tested.
In one embodiment of the present invention, a multi-point probe is used. In a particular embodiment, the multi-point probe is a 12-point probe, although probes having higher or lower number of points are also useful. This 12-point probe is a micro-machined probe and there are a variety of spacings between the probes. Moreover, current and voltage can be routed to any four probes of the 12-point probe. Using the twelve-point probe, a large array of resistance data can be quickly determined. With this multi-probe, it is more efficient to determine resistance data for a variety of probe spacings, then apply a magnetic field to switch magnetization of the free layer, and again determine resistance data for a variety of probe spacings.
In step 710, measured in-plane MR is determined for each set of high and low resistances for each spacing. For instance, if a probe has a spacing of 10 microns between each of the probes in a four-point probe, a resultant high and low resistance can be determined for this 10 micron spacing. From the high and low resistances, a corresponding in-plane MR is determined for this 10 micron spacing.
In step 730, curve fitting is performed. The RA product, RF, RP, and perpendicular MR are determined by using one set of determined resistances from either the high or low resistance data and by using a curve fitting technique. In other words, a mathematical equation has already been determined for the particular configuration of probes or contact pads. Fitting is performed by assuming values for the RA product, RF, RP, and perpendicular MR and then calculating in-plane MR and resistance curves. The values of the in-plane MR and resistance curves are then compared with the calculated values of in-plane MR and resistance curves. This process is then iterated, each time changing values for the RA product, RF, RP, and perpendicular MR, until the best agreement between the measured and calculated values of in-plane MR and resistance is obtained.
Thus, method 700 allows perpendicular MR, the RA product, RF, and RP to be determined with little or no processing, in contrast to conventional techniques, which required lengthy processing.
It should be noted that the “curves” described above should be broadly construed. For instance, a 12-point probe is described below that allows various spacings to be used during measurements. Because the spaces are not evenly chosen, all of the resultant in-plane MR and resistance data cannot be shown on a graph having two axis. Nonetheless, curve fitting techniques and theoretically derived equations are available to fit the equations with the data.
Advantageously, steps 710 and 720 of method 700 shown in
Turning now to
Referring now to
The tunnel junction test device 900 comprises a magnetic field generator 910, multi-point probe 915, multiplexer 930, resistance module 958, control module 955, and data storage module 960. Resistance module 958 comprises a current module 940 and a voltage module 950. Current module 940 measures current, such that the current module 940 produces or determines or both a current to within a particular accuracy. Voltage module 950 measures voltage, such that the voltage module 950 produces or determines or both a voltage to within a particular accuracy. Similarly, The connections 926-1 through 926-N (collectively, “connections” 926) connect the multi-point probe 915 and the multiplexer 930. Current module 940 is coupled to multiplexer 930 through current lines 941 and 942. The voltage module 950 is similarly coupled to the multiplexer 930 through voltage lines 951 and 952. Connection 953 is an optional connection used to communicate the measured resistance to data storage 960, if used.
Computer system 970 comprises memory 973 having an analysis module 980, a processor 975, and a media interface 977. Computer system 970 is shown interacting with a Digital Versatile Disk (DVD) 978 through media interface 977. Analysis module 980 performs the method 700 of
Advantageously, the smallest spacing between any two of the multiple probes on multi-point probe 915 used during a resistance measurement has a spacing of 100 microns or less. In other words, multi-point probe 915 comprises four or more probes. Two of the probes used during a resistance measurement will have the smallest spacing of the four probes being used during a resistance measurement. In order to characterize a large number of tunnel junction film stacks of different materials, this smallest spacing is beneficially 100 microns or less.
As is known in the art, the methods and apparatus discussed herein may be distributed as an article of manufacture that itself comprises a computer-readable medium having computer-readable code means embodied thereon. The computer-readable program code means is operable, in conjunction with a computer system such as computer system 970, to carry out all or some of the steps to perform the methods or create the apparatuses discussed herein. The computer-readable medium may be a recordable medium (e.g., floppy disks, hard drives, compact disks, DVD 978, or memory cards) or may be a transmission medium (e.g., a network comprising fiber-optics, the world-wide web, cables, or a wireless channel using time-division multiple access, code-division multiple access, or other radio-frequency channel). Any medium known or developed that can store information suitable for use with a computer system may be used. The computer-readable code means is any mechanism for allowing a computer to read instructions and data, such as magnetic variations on a magnetic medium or height variations on the surface of an optical disk, such as DVD 978.
Memory 973 configures the processor 975 to implement the methods, steps, and functions disclosed herein. The memory 973 could be distributed or local and the processor 975 could be distributed or singular. The memory 973 could be implemented as an electrical, magnetic or optical memory, or any combination of these or other types of storage devices. Moreover, the term “memory” should be construed broadly enough to encompass any information able to be read from or written to an address in the addressable space accessed by processor 973. With this definition, information on a network (not shown) is still within memory 973 because the processor 975 can retrieve the information from the network. It should be noted that each distributed processor that makes up processor 973 generally contains its own addressable memory space. It should also be noted that some or all of computer system 970 can be incorporated into an application-specific or general-use integrated circuit.
It should be noted that a tunnel junction test device 900 will generally comprise additional devices that are not shown. For example, because the probes and contact pads are usually very small, some type of magnification is beneficial. This magnification may be provided by a Charge-Coupled Device (CCD). Additionally, a mounting system is generally used to mount the multi-point probe 915 and semiconductor wafer 920 and to move each relative to the other (or to keep the semiconductor wafer 920 fixed and move the multi-point probe 915, for instance). Specifically, a mechanism for positioning probes can use a laser beam that reflects off the bottom of the semiconductor wafer 920. It is also beneficial to be able to measure multiple locations on the semiconductor wafer 920, in order to determine uniformity of the stack and its associated layers.
Various configurations of apparatus 990 will now be described in reference to different probes and contact pads.
The largest and smallest spacings between selected probes are chosen to be within a predetermined amount from the expected length scale, λ, based on the expected RA product, RF and RP. The spacing which is used to determine the predetermined amount is an average of the three spacings between the four selected probes. Because the micro-machined probe 1000 is formed through known semiconductor micro-machining techniques, generally the smallest possible spacing that can be made will be used in order to come as close as possible to determining RF. Currently, spacings suitable to measure an RA product of 200 Ohm-micron2 have been created, which is a spacing as small as 1.5 microns, depending on RF and RP. Advantageously, the smallest spacing between two of the plurality of probes used during a resistant measurement is a spacing of approximately 100 microns or less, as this provides suitable spacings to measure tunnel junction stack films with a wide variety of RA products, and resistances per square. In the example of
Probes 1001 through 1012 are placed in physical contact with the surface 925 of semiconductor wafer 920 (in this example, contact pads 923 are not used), which also provides an electrical connection between the probes 1001 through 1012 and the stack on the semiconductor wafer 920. The magnetization of the stack is set in a first magnetization by using the magnetic field generator 910. Control module 955 selects appropriate probes 1001 through 1012 in order to route current and voltage through four probes of the 12-point probe 1000, and control module 955 controls resistant module 958 so that an appropriate resistance is measured. This process is repeated for a predetermined number of combinations of probes 1001 through 1012. Resultant resistance data is stored in data storage 960. Then magnetic field generator 910 is turned on by the control module 955, and the process of selecting four of the probes 1001 through 1012 of 12-point probe 1000 is repeated. The resultant resistance data 953 is stored in data storage 960. The analysis module 980 then retrieves the probe data 961 and analyzes the data to create the perpendicular MR, the RA product, and the resistances RF and RP.
Using the micro-machined 12-point probe 1000 in tunnel junction test device 900 provides very fast voltage determination times with no required stack processing.
Referring now to
The four-point probe 1055 is used similarly to the 12-point probe 1000. The four-point probe 1055 is placed in contact with the surface 925 of semiconductor wafer 920. Current is passed through probes 1061 and 1064 and voltage is measured by using probes 1062 and 1063. Alternatively, voltage is placed on probes 1062 and 1063 and current is measured from probes 1061 and 1064. The magnetic field generator 910 is turned on, which switches the magnetization of the free layer in the stack. A measurement of resistance is then performed. Because the four-point probe 1055 is designed for one spacing, multiple four-point probes 1055 with multiple spacings need to be used to provide meaningful data in order for analysis module 980 to characterize tunnel junction quantities. The spacings are provided from as small as possible to a predetermined distance from λ. The space a is used to determine the predetermined distance. Generally, each four-point probe is manually placed in contact with the semiconductor wafer 920. It is beneficial that the smallest spacing between two of the plurality of probes used during a resistant measurement is a spacing of approximately 100 microns or less.
Referring now to
The equations for the configuration of
Turning now to
Generally, a conventional four-point probe is used to physically contact the pads 1235 of four-point contact pads 1230. Generally, the multiplexer 930 will not be used and, instead a four-point probe is connected to the pads 1235. Current is passed through pads 1235-1 and 1235-4 and voltage is measured across pads 1235-2 and 1235-4. Multiple different four-point contact pads 1230 will be formed during processing, each with different spacing L. Data from each of these contact pads 1230 will be used to characterize the stack 1210.
Similarly, the other configurations shown in
Referring now to
Referring now to
It should be noted that the stack shown in the preceding figures may be much more complex than that described. In particular, many such stacks have free and pinned layers are themselves made of many layers. Additionally, the free and pinned layers may be reversed, such that the free layer is on top of the substrate and the pinned layer then is used to perform characterization measurements. There are also materials having more than two magnetizations, and tunnel junction film stacks where the electrodes (i.e., the layers above and below the tunnel junction barrier) are nonmagnetic and the tunnel junction barrier is magnetic. Moreover, there are tunnel junction film stacks having multiple tunnel junction barriers, free layers, and pinned layers. The present invention can be used with any tunnel junction stack films.
It is to be understood that the embodiments and variations shown and described herein are merely illustrative of the principles of this invention and that various modifications may be implemented by those skilled in the art without departing from the scope and spirit of the invention.